There are a couple of things I would just like to mention in passing. Jenny Ruhl has just posted a nice entry about ACCORD. This is very important. Lowering your blood glucose is significantly protective against CVD events. This is the exact opposite of the initial analysis of the results where a flawed interpretation of the data led to vociferous suggestions that lowering the HbA1c of diabetics might be actually dangerous. Hopefully this reanalysis will put an end to such stupid ideas which are still dangerously prevalent today.
Before Jenny posted the above I had been thinking about both metformin and insulin for the management of diabetes. I have posted on insulin, which is probably the ideal drug for diabetes management provided it is combined with low carbohydrate eating, in the past but it bears reiterating.
This is my opinion. If you can control your diabetes with metformin and wish to eat lots of carbohydrate, by all means get on with it, that's your choice. Not checking or not worrying about your blood glucose excursions might be a mistake. What you mean by good control might not involve HbA1c values in the 5% region or below.
If you need insulin to control your blood glucose, you have no choice. It's low carb. Live with it.
Some researchers (who write third rate papers using a rather inappropriate pancreatectomised dog model) are now starting to wake up to the fact that the pancreas secretes insulin in to the portal vein, not the subcutaneous tissues. In normal individuals insulin/glucose arrives at the liver via the portal vein and insulin facilitates transport of the glucose in to the liver, being metabolised in the process. Relatively little post pranial insulin or glucose penetrates to the peripheral circulation in a normal individual.
EDIT: Gretchen has kindly pointed out that the liver uptakes glucose through GLUT2, not GLUT4. The function of insulin, acting on its receptor (facilitating its degradation), is to suppress gluconeogenesis and hepatic glucose output. This is the correct function of the elevated portal insulin level. It makes no difference to the issues with peripheral vs portal insulin, but the correction is welcome. END EDIT
Once a person needs insulin to control their blood sugar levels they inject it subcutaneously. This will invariably elevate the systemic concentration of insulin. It will only modestly elevate the portal vein level. This is very important.
In the cited paper the dogs get a meal with 50% of calories from carbohydrate, 30% from fat and 20% from protein. In control dogs, instrumented but not pancreatectomised, the portal vein insulin after the meal is, at certain time points, ten times the peripheral systemic concentration. This is what is needed to allow the liver deal with the glucose load from the meal while simultaneously protecting the body from both hyperglycaemia and hyperinsulinaemia.
Subcutaneous insulin will make the rest of the body do the liver's job of clearing post prandial glucose, the liver can't manage because it never sees the requisite ten times the normal peripheral insulin concentration needed to deal effectively with the portal glucose load.
What happens when you use the rest of the body as a glucose sump? From the paper:
"Peripheral hyperinsulinemia is associated not only with increased risk of hypoglycemia, but also an increase in catechol and cortisol secretion and lipolysis , deleterious effects on vessel walls , ischemic heart disease, hypertension and hyperlipidemia , and abnormalities in hemostasis "
As always, I would add that it's the obese, blind, legless person in the queue for dialysis who pays the bill for eating the carbohydrate.
This is the situation for all type I diabetics and the more advanced type 2 diabetics. The only route round it is to use intra peritoneal insulin which is, in part, absorbed through the mesenteric veins so is partially portal vein selective. There are, needless to say, a stack of complications to intra peritoneal insulin infusion. Tight control of glucose using subcutaneous insulin from a blood glucose controlled pump is no solution. Though glycaemia is better controlled it is still at the cost of too little insulin in the portal vein and too much in the periphery, using the body as a glucose sump. Over the years I have never been quite able to decide whether hyperinsulinaemia or hyperglycaemia is the primary factor which kills nerves and kidneys. It's a difficult call. And a fascinating discussion in its own right.
What happens if you eat a diet very low in starch?
Very little insulin is ever secreted by the pancreas, especially as glucokinase down regulates. Very little glucose ever needs to be taken up by the liver. Very little insulin will be metabolised by the liver. The insulin gradient between the portal vein and the systemic circulation will be as low as you can practically get it. If someone still needs to inject insulin alongside a very low carbohydrate diet, and many might not, injecting a very small amount subcutaneously will deliver an arterial concentration to the gut, pancreas and eventually to the portal vein and liver which is still quite close to what the portal vein might have supplied. If the body is not using insulin the tissues will not extract it, so portal and systemic concentrations will converge. Everything pans out at some near basal level.
A very low carbohydrate diet is not perfect for insulin dependent diabetics but it is streets ahead of anything else. What people do or do not consider a "normal" human diet will not get around this. Need exogenous insulin? You are not in a position to eat ancestral starch. It's a simple matter of anatomy, physiology and biochemistry.